Abnormal transformer area diagnosis method and device

Abstract

The invention discloses an abnormal transformer area diagnosis method and device, and the method comprises the steps: 1, collecting the power consumption and power supply of a transformer area, calculating the predicted power consumption and residual error of a diagnosis day according to the power consumption and power supply of the transformer area, and judging whether the power supply or power consumption of the day is abnormal or not according to the residual error; step 2, if the power supply quantity or the power consumption quantity of the day is judged to be abnormal in the step 1, calculating abnormal loss-reducible power quantity, and trying to calculate the loss-reducible power quantity and the line loss rate of the transformer area based on the abnormal loss-reducible power quantity; and step 3, outputting a calculation result in the step 2 in a list mode. According to the method, whether the power supply quantity or the electricity consumption is abnormal or not is judged by analyzing the electricity consumption and the power supply quantity, and when the electricity consumption is abnormal, the transformer area abnormality diagnosis is realized through the loss-reducible electricity quantity, the trial calculation transformer area loss-reducible electricity quantity and the line loss rate so as to maintain the stability of the transformer area.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormality diagnosis of substations, and in particular to a method and device for diagnosing abnormal substations. Background Art

[0002] Substation power supply is an important part of the power system. It is responsible for reducing the voltage of the electric energy transmitted by the high-voltage transmission line and supplying it to the end users. With the iterative development of technology, substation power supply is gradually being managed intelligently, and the introduction of smart energy meters and remote monitoring systems has improved the efficiency of power supply management. It is necessary to analyze the power supply and power consumption data of the substation and perform abnormal diagnosis to maintain the stability of the substation. Summary of the invention

[0003] In view of the technical defects and technical drawbacks in the prior art, the embodiments of the present invention provide an abnormal area diagnosis method and device that overcomes the above problems or at least partially solves the above problems. The specific scheme is as follows:

[0004] As a first aspect of the present invention, a method for diagnosing an abnormal station area is provided, the method comprising:

[0005] Step 1: Collect the power consumption and power supply of the substation area, calculate the predicted power consumption and residual on the diagnosis day based on the power consumption and power supply of the substation area, and determine whether the power supply or power consumption on that day is abnormal based on the residual;

[0006] Step 2: If the power supply or power consumption on that day is determined to be abnormal in step 1, the abnormal amount of power that can be reduced is calculated, and based on the abnormal amount of power that can be reduced, the amount of power that can be reduced and the line loss rate of the substation are calculated;

[0007] Step 3: Output the calculation results of step 2 in the form of a list.

[0008] Furthermore, step 1 specifically includes:

[0009] A clustering algorithm is used to cluster the historical power supply and power consumption of the substation for a certain period of time, and a linear fit is performed on the historical power supply and power consumption after clustering to obtain a corresponding fitting curve model. The power supply on the diagnosis day is substituted into the fitting curve model to obtain the predicted power consumption on the diagnosis day, and the residual is calculated based on the predicted power consumption and the actual power consumption, wherein the residual = actual power consumption - predicted power consumption, and the residual is the amount of electricity that can be reduced; when the absolute value of the residual is greater than the set value K, it is judged that the power supply or power consumption on that day is abnormal, and the corresponding amount of electricity that can be reduced is the abnormal amount of electricity that can be reduced.

[0010] Furthermore, in step 2, the power reduction amount of the substation = the sum of all abnormal power reduction amounts. When the residual of the power reduction amount of the substation and the linear fitting does not exceed the predetermined value Q, the corresponding model is considered to be accurate, that is, the residual calculated by the substation fitting is close to the sum of all abnormal power reduction amounts, indicating that the two results confirm each other and the substation is suitable for the model theory.

[0011] During the line loss rate calculation process, when the abnormality belongs to the terminal, the line loss rate = (predicted power supply - power consumption + abnormal power loss reduction of the corresponding terminal) / predicted power supply * 100%.

[0012] Furthermore, the method further includes: determining the source of the abnormality according to the power consumption and power supply 7 days before the diagnosis date, specifically including:

[0013] Collect the power consumption and power supply 7 days before the diagnosis day, and construct the power consumption array and power supply array respectively;

[0014] Calculate the standard deviations x and y of the power consumption array and the power supply array respectively;

[0015] When x>K1*y, the anomaly comes from the user in the substation area;

[0016] When y>K2*x, the anomaly comes from the total table of the station area;

[0017] Among them, k1 is a preset first threshold, and k2 is a preset second threshold.

[0018] Furthermore, the method further comprises:

[0019] Constructing sample data, wherein the sample data is derived from historical cases, including abnormal terminal metering data, abnormal terminal collection data, abnormal user metering data, abnormal user collection data, abnormal archive data, and breach of contract electricity consumption data;

[0020] Constructing a decision tree model: including constructing dynamic and static features of the decision tree model, where the static features are archival data and the dynamic features are line loss related features; constructing statistical features for the dynamic features, and combining windowing to construct derivative features of the dynamic features;

[0021] The main cause of the abnormality in the substation area is determined through the constructed decision tree model.

[0022] Furthermore, the method further comprises: selecting 80% of the sample data from the sample data as expert samples, and selecting 20% ​​of the sample data as test samples.

[0023] Furthermore, the method further includes: performing a diagnosis of abnormal data users in the substation area, specifically including:

[0024] Based on user profiles, electricity consumption and load data, data anomaly diagnosis is performed through consistency, contradiction, missingness, extreme values ​​and repeatability to diagnose abnormal users and abnormal time periods.

[0025] As a second aspect of the present invention, there is provided an abnormal area diagnosis device, the device comprising:

[0026] The abnormality judgment module is used to collect the power consumption and power supply of the substation area, calculate the predicted power consumption and residual of the diagnosis day according to the power consumption and power supply of the substation area, and judge whether the power supply or power consumption of the day is abnormal according to the residual;

[0027] A calculation module is used to calculate the abnormal amount of power that can be reduced when it is judged that the power supply or power consumption on that day is abnormal, and based on the abnormal amount of power that can be reduced, calculate the amount of power that can be reduced and the line loss rate of the substation area;

[0028] The output module is used to output the calculation results of the calculation module in the form of a list.

[0029] Furthermore, the abnormality judgment module is specifically used for:

[0030] A clustering algorithm is used to cluster the power supply and power consumption of the substation, and then a linear fit is performed. The predicted power consumption and residual of the diagnosis day are calculated based on the power supply, where the residual = actual power consumption - predicted power consumption, and the residual is the amount of electricity that can be reduced. When the absolute value of the residual is greater than the set value K, it is judged that the power supply or power consumption on that day is abnormal, and the corresponding amount of electricity that can be reduced is the abnormal amount of electricity that can be reduced.

[0031] Furthermore, the calculation of the power reduction amount and line loss rate of the area based on the abnormal power reduction amount is specifically as follows:

[0032] The power loss reduction in the substation area = the sum of all abnormal power loss reduction. When the residual of the power loss reduction in the substation area and the linear fitting does not exceed K, the assumption is valid;

[0033] During the line loss rate calculation process, when the abnormality belongs to the terminal, the line loss rate = (predicted power supply - power consumption + abnormal power loss reduction of the corresponding terminal) / predicted power supply * 100%.

[0034] The present invention has the following beneficial effects:

[0035] The present invention is based on the marketing system and the electricity consumption information collection system, integrates multi-source data, takes "full analysis, precise positioning, and hierarchical recommendation" as the goal, uses big data and machine learning algorithms, and focuses on archives, collection, measurement, technology, electricity theft and other dimensions. Through the quantification of line loss anomalies and impact analysis, combined with the line loss characteristics of the substation, a hierarchical recommendation strategy for substation line loss management is formed to solve the problems of complex and difficult positioning of line loss causes, huge data volume and difficulty in analysis, and weak basic capabilities and high pressure. It realizes hierarchical management recommendations for abnormal substations, improves the efficiency of line loss rectification, and reduces the burden on the grassroots. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of a flow chart of an abnormal area diagnosis method provided by an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of the framework of an abnormal station area diagnosis device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] like Figure 1 As shown, the present invention provides an abnormal area diagnosis method, comprising the following steps:

[0040] Step 1: Collect the power consumption and power supply of the substation area, calculate the predicted power consumption and residual on the diagnosis day based on the power consumption and power supply of the substation area, and determine whether the power supply or power consumption on that day is abnormal based on the residual;

[0041] Step 2: If the power supply or power consumption on that day is determined to be abnormal in step 1, the abnormal amount of power that can be reduced is calculated, and based on the abnormal amount of power that can be reduced, the amount of power that can be reduced and the line loss rate of the substation are calculated;

[0042] Step 3: Output the calculation results of step 2 in the form of a list.

[0043] The present invention determines whether the power supply or power consumption is abnormal by analyzing the power consumption and power supply. When an abnormality occurs, the present invention calculates the power loss reduction and line loss rate of the substation through the power loss reduction, thereby realizing the abnormal diagnosis of the substation to maintain the stability of the substation.

[0044] In some embodiments, step 1 specifically includes:

[0045] A clustering algorithm is used to cluster the historical power supply and power consumption of the substation for a certain period of time, and a linear fit is performed on the historical power supply and power consumption after clustering to obtain a corresponding fitting curve model. The power supply on the diagnosis day is substituted into the fitting curve model to obtain the predicted power consumption on the diagnosis day, and the residual is calculated based on the predicted power consumption and the actual power consumption, wherein the residual = actual power consumption - predicted power consumption, and the residual is the amount of electricity that can be reduced; when the absolute value of the residual is greater than the set value K, it is judged that the power supply or power consumption on that day is abnormal, and the corresponding amount of electricity that can be reduced is the abnormal amount of electricity that can be reduced.

[0046] In some embodiments, step 2 comprises:

[0047] The power loss reduction in the substation area = the sum of all abnormal power loss reduction. When the residual of the power loss reduction in the substation area and the linear fitting does not exceed K, the assumption is valid;

[0048] During the line loss rate calculation process, when the abnormality belongs to the terminal, the line loss rate = (predicted power supply - power consumption + abnormal power loss reduction of the corresponding terminal) / predicted power supply * 100%.

[0049] In some embodiments, the method further includes determining the source of the abnormality, specifically including:

[0050] Collect the power consumption and power supply 7 days before the diagnosis day, and construct the power consumption array and power supply array respectively;

[0051] Calculate the standard deviations x and y of the power consumption array and the power supply array respectively;

[0052] When x>K1*y, the anomaly comes from the user in the substation area;

[0053] When y>K2*x, the anomaly comes from the total table of the station area;

[0054] Wherein, k1 is a preset first threshold, k2 is a preset second threshold, and in this embodiment, k1=1.1 and k2=1.2 are preset.

[0055] In some specific embodiments, the method further includes: weak classification to determine the main cause of the station area, and the steps of weak classification to determine the main cause of the station area are as follows:

[0056] a. Construct sample data, select 80% of the samples as expert samples and 20% as test samples; the sample data comes from historical cases: abnormal terminal metering data, abnormal terminal collection data, abnormal user metering data, abnormal user collection data, abnormal archive data and breach of contract electricity consumption data; all data include abnormal data and normal data

[0057] b. Construct features and machine learning. When constructing features and machine learning, the features are divided into static attributes and dynamic features. The static attributes are the substation archive information, and the dynamic features are line loss related features. Statistical features are constructed for dynamic features, and derived features of dynamic features are constructed by combining windowing.

[0058] c. Determine the main cause of the abnormality in the substation through the constructed decision tree model.

[0059] In some specific embodiments, the method further includes: user diagnosis of abnormal data in the station area:

[0060] The abnormal data diagnosis of the substation is based on the archives, electricity consumption, load and other data. The data abnormality diagnosis is performed through consistency, contradiction, missingness, extreme value and repeatability to diagnose abnormal users and abnormal time periods, improve the accuracy of abnormal diagnosis analysis, and perform correlation calculation based on abnormal data and the main cause residual of the substation. The correlation calculation is mainly carried out by constructing vectors and calculating vector values. Whether it includes single abnormality and loss-reducing electricity, diagnose the abnormal electricity data, and the abnormal data includes abnormal users, abnormal time and single abnormality and loss-reducing electricity.

[0061] In some embodiments, when calculating the amount of electricity that can be reduced, the user's daily electricity consumption and the power supply of the substation are supplemented according to the abnormal electricity data. The amount of electricity that can be reduced for a single abnormality = predicted electricity - electricity recorded by the system.

[0062] like Figure 2 As shown, the present invention provides an abnormal area diagnosis device, comprising:

[0063] The abnormality judgment module is used to collect the power consumption and power supply of the substation area, calculate the predicted power consumption and residual of the diagnosis day according to the power consumption and power supply of the substation area, and judge whether the power supply or power consumption of the day is abnormal according to the residual;

[0064] A calculation module is used to calculate the abnormal amount of power that can be reduced when it is judged that the power supply or power consumption on that day is abnormal, and based on the abnormal amount of power that can be reduced, calculate the amount of power that can be reduced and the line loss rate of the substation area;

[0065] The output module is used to output the calculation results of the calculation module in the form of a list.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for diagnosing abnormal areas, characterized in that: The method comprises: Step 1: Collect the power consumption and power supply of the substation area, calculate the predicted power consumption and residual on the diagnosis day based on the power consumption and power supply of the substation area, and determine whether the power supply or power consumption on that day is abnormal based on the residual; Step 2: If the power supply or power consumption on that day is determined to be abnormal in step 1, the abnormal amount of power that can be reduced is calculated, and based on the abnormal amount of power that can be reduced, the amount of power that can be reduced and the line loss rate of the substation are calculated; Step 3: Output the calculation results of step 2 in the form of a list.

2. The abnormal station area diagnosis method according to claim 1, characterized in that: Step 1 specifically includes: A clustering algorithm is used to cluster the historical power supply and power consumption of the substation for a certain period of time, and a linear fit is performed on the historical power supply and power consumption after clustering to obtain a corresponding fitting curve model. The power supply on the diagnosis day is substituted into the fitting curve model to obtain the predicted power consumption on the diagnosis day, and the residual is calculated based on the predicted power consumption and the actual power consumption, wherein the residual = actual power consumption - predicted power consumption, and the residual is the amount of electricity that can be reduced; when the absolute value of the residual is greater than the set value K, it is judged that the power supply or power consumption on that day is abnormal, and the corresponding amount of electricity that can be reduced is the abnormal amount of electricity that can be reduced.

3. The abnormal station area diagnosis method according to claim 2, characterized in that: In step 2, the power reduction amount in the substation = the sum of all abnormal power reduction amounts. When the residual error between the power reduction amount in the substation and the linear fitting does not exceed the predetermined value Q, the model is considered accurate. During the line loss rate calculation process, when the abnormality belongs to the terminal, the line loss rate = (predicted power supply - power consumption + abnormal power loss reduction of the corresponding terminal) / predicted power supply * 100%.

4. The abnormal station area diagnosis method according to claim 1, characterized in that: The method further includes: determining the source of the abnormality based on the power consumption and power supply 7 days before the diagnosis date, specifically including: Collect the power consumption and power supply 7 days before the diagnosis day, and construct the power consumption array and power supply array respectively; Calculate the standard deviations x and y of the power consumption array and the power supply array respectively; When x>K1*y, the anomaly comes from the user in the substation area; When y>K2*x, the anomaly comes from the total table of the station area; Among them, k1 is a preset first threshold, and k2 is a preset second threshold.

5. The abnormal station area diagnosis method according to claim 1, characterized in that: The method further comprises: Constructing sample data, wherein the sample data is derived from historical cases, including abnormal terminal metering data, abnormal terminal collection data, abnormal user metering data, abnormal user collection data, abnormal archive data, and breach of contract electricity consumption data; Constructing a decision tree model: including constructing dynamic and static features of the decision tree model, where the static features are archival data and the dynamic features are line loss related features; constructing statistical features for the dynamic features, and combining windowing to construct derivative features of the dynamic features; The main cause of the abnormality in the substation area is determined through the constructed decision tree model.

6. The abnormal station area diagnosis method according to claim 5, characterized in that: The method further comprises: selecting 80% of the sample data from the sample data as expert samples, and selecting 20% ​​of the sample data as test samples.

7. The abnormal station area diagnosis method according to claim 1, characterized in that: The method further includes: performing a diagnosis of abnormal data users in the substation area, specifically including: Based on user profiles, electricity consumption and load data, data anomaly diagnosis is performed through consistency, contradiction, missingness, extreme values ​​and repeatability to diagnose abnormal users and abnormal time periods.

8. An abnormal area diagnosis device, characterized in that: The device comprises: The abnormality judgment module is used to collect the power consumption and power supply of the substation area, calculate the predicted power consumption and residual of the diagnosis day according to the power consumption and power supply of the substation area, and judge whether the power supply or power consumption of the day is abnormal according to the residual; A calculation module is used to calculate the abnormal amount of power that can be reduced when it is judged that the power supply or power consumption on that day is abnormal, and based on the abnormal amount of power that can be reduced, calculate the amount of power that can be reduced and the line loss rate of the substation area; The output module is used to output the calculation results of the calculation module in the form of a list.

9. The abnormal station area diagnosis device according to claim 8, characterized in that: The abnormality judgment module is specifically used for: A clustering algorithm is used to cluster the power supply and power consumption of the substation, and then a linear fit is performed. The predicted power consumption and residual of the diagnosis day are calculated based on the power supply, where the residual = actual power consumption - predicted power consumption, and the residual is the amount of electricity that can be reduced. When the absolute value of the residual is greater than the set value K, it is judged that the power supply or power consumption on that day is abnormal, and the corresponding amount of electricity that can be reduced is the abnormal amount of electricity that can be reduced.

10. The abnormal station area diagnosis device according to claim 9, characterized in that: The above-mentioned abnormally reduced power consumption and the calculated power consumption and line loss rate of the substation area are specifically as follows: The power loss reduction in the substation area = the sum of all abnormal power loss reduction. When the residual error between the power loss reduction in the substation area and the linear fitting does not exceed K, the assumption is valid. During the line loss rate calculation process, when the abnormality belongs to the terminal, the line loss rate = (predicted power supply - power consumption + abnormal power loss reduction of the corresponding terminal) / predicted power supply * 100%.

Images